Flexible Manufacturing Systems of Unreliable Machines

Abstract

In this chapter we study Markovian models for the performance evaluation of Flexible Manufacturing Systems (FMSs). In recent years there has been an increasing role of computers in manufacturing. One important area of Computer Aided Manufacturing (CAM) is FMS. The advantage of an FMS is that it can reduce the in the system and increases the machine utilization when suitable production policy is implemented. Moreover, it can also reduce manufacturing lead time and labor cost. However, the setting up cost and the maintenance cost of machines are high in an FMS; see Buzacott and Shanthikumar [17] for instance. Due to the high capital investment, an FMS is only considered to operate economically if there is a high level of system performance. The system design and management are therefore important front considerations for setting up an FMS. Mathematical modeling can help with decisions required to design and manage an FMS. Queuing theory is a useful tool for modeling manufacturing systems; see for instance [44, 46, 110]. In fact, most analytical models describe an FMS as a queuing system, in which the customers are jobs to be processed or product in inventory and the servers are simply the reliable machines (workstations) in the system; see for instance Buzacott and Yao [18]. However, the assumption that the machines are reliable can greatly affect the performance evaluation of an FMS and therefore should be taken into account in any proposed model.